FlexPS: Flexible Parallelism Control in Parameter Server Architecture
Summary: FlexPS extends parameter-server systems with a multi-stage abstraction enabling runtime-adjustable parallelism for dynamic ML workloads. Its stage scheduler, stage-aware consistency, and direct model transfer substantially improve execution over existing PS systems. (summarized by gpt-5.6-luna on Jul 24 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Yuzhen Huang (Chinese University of Hong Kong)
- 2. Tatiana Jin (Chinese University of Hong Kong)
- 3. Yidi Wu (Chinese University of Hong Kong)
- 4. Zhenkun Cai (Chinese University of Hong Kong)
- 5. Xiao Yan (Chinese University of Hong Kong)
- 6. Fan Yang (Chinese University of Hong Kong)
- 7. Jinfeng Li (Chinese University of Hong Kong)
- 8. Yuying Guo (Chinese University of Hong Kong)
- 9. James Cheng (Chinese University of Hong Kong)
BibTeX Citation
@article{huang_vldb18,
title = {{FlexPS: Flexible Parallelism Control in Parameter Server Architecture}},
author = {Huang, Yuzhen and Jin, Tatiana and Wu, Yidi and Cai, Zhenkun and Yan, Xiao and Yang, Fan and Li, Jinfeng and Guo, Yuying and Cheng, James},
journal = {PVLDB},
series = {{VLDB} '18},
volume = {11},
number = {5},
pages = {566--579},
doi = {10.14778/3172077.3172085},
url = {https://doi.org/10.14778/3172077.3172085},
year = {2018}
}Incoming Citations (Sorted by Pagerank)
Showing 8 of 8 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 1,152 | Cerebro: A Data System for Optimized Deep Learning Model Selection | 2020 | VLDB | 0.00011796404 |
| 3,242 | Towards Demystifying Serverless Machine Learning Training | 2021 | SIGMOD | 7.4967268e-05 |
| 3,766 | CROSSBOW: Scaling Deep Learning with Small Batch Sizes on Multi-GPU Servers | 2019 | VLDB | 7.0360981e-05 |
| 4,097 | Distributed Deep Learning on Data Systems: A Comparative Analysis of Approaches | 2021 | VLDB | 6.8064316e-05 |
| 5,635 | NuPS: A Parameter Server for Machine Learning with Non-Uniform Parameter Access | 2022 | SIGMOD | 6.0549984e-05 |
| 5,999 | BAGUA: Scaling up Distributed Learning with System Relaxations | 2022 | VLDB | 5.9170896e-05 |
| 6,644 | Dynamic Parameter Allocation in Parameter Servers | 2020 | VLDB | 5.7224864e-05 |
| 8,495 | Just Move It! Dynamic Parameter Allocation in Action | 2021 | VLDB | 5.3292461e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 8 of 8 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 22 | Distributed GraphLab: A Framework for Machine Learning and Data Mining in the Cloud | 2012 | VLDB | 0.00055938421 |
| 464 | An Architecture for Parallel Topic Models | 2010 | VLDB | 0.00017783871 |
| 2,186 | Heterogeneity-aware Distributed Parameter Servers | 2017 | SIGMOD | 8.8916253e-05 |
| 2,667 | NOMAD: Non-locking, stOchastic Multi-machine algorithm for Asynchronous and Decentralized matrix completion | 2014 | VLDB | 8.151622e-05 |
| 3,863 | Husky: Towards a More Efficient and Expressive Distributed Computing Framework | 2016 | VLDB | 6.961908e-05 |
| 4,902 | Scalable Asynchronous Gradient Descent Optimization for Out-of-Core Models | 2017 | VLDB | 6.3617262e-05 |
| 8,445 | LFTF: A Framework for Efficient Tensor Analytics at Scale | 2017 | VLDB | 5.3324907e-05 |
| 12,291 | The Best of Both Worlds: Big Data Programming with Both Productivity and Performance | 2017 | SIGMOD | 4.9769913e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 8,426 | Demonstrating PDSP-Bench: A Benchmarking System for Parallel and Distributed Stream Processing | 2025 | SIGMOD |
| 2 | 10,904 | Resilience-Aware Elastic Scaling for Cloud-Native Online DL Training on Multi-Tenant GPU Clusters | 2026 | VLDB |
| 3 | 7,317 | Model-Free Control for Distributed Stream Data Processing using Deep Reinforcement Learning | 2018 | VLDB |
| 4 | 2,247 | PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallel | 2023 | VLDB |
| 5 | 1,082 | Hybrid Parallelization Strategies for Large-Scale Machine Learning in SystemML | 2014 | VLDB |
| 6 | 8,495 | Just Move It! Dynamic Parameter Allocation in Action | 2021 | VLDB |
| 7 | 2,186 | Heterogeneity-aware Distributed Parameter Servers | 2017 | SIGMOD |
| 8 | 6,644 | Dynamic Parameter Allocation in Parameter Servers | 2020 | VLDB |
| 9 | 5,437 | PS2: Parameter Server on Spark | 2019 | SIGMOD |
| 10 | 5,635 | NuPS: A Parameter Server for Machine Learning with Non-Uniform Parameter Access | 2022 | SIGMOD |